Red Sox pitcher Curt Schilling last March correctly predicted, in theory, the winner of the American League East when he said, "The rotation that makes the most starts wins the division. It's that simple." Boston's projected five-man rotation made 140 starts, tops in the division; the Red Sox won the division. The Yankees' season-opening rotation made 105 starts.
In 2006 it was New York that won the battle over Boston for most starts by its top five starters, 125-107 -- and also won the division.
The Schilling Theory got me thinking that its applications go beyond the Boston-New York rivalry. As players are better trained than ever before and as front offices make better use of available information than ever before, the difference between going home or to the playoffs may rest more on rotation stability than any other factor -- which means pennant races might well be decided by the happenstance of injury. In other words, you can pour enormous resources and planning into team building and yet you're left to the mercy of when the alarm clocks that are ulnar collateral nerves or rotator cuffs decide to go off.
In coming weeks I'll get to what this means for the 2008 Red Sox and Yankees -- and the seven most at-risk young starters in baseball -- but let's just say for now that New York would be much better off in the short run, though not as certainly in the long run, to trade for Twins ace Johan Santana before spring training. The Yankees can build years of success around young starting pitchers Joba Chamberlain, Phil Hughes and Ian Kennedy, but just not this year without some risk. The Yankees know they can't push any of those starters to 200 innings this year -- not at their ages and not given the risky leap in workload it would require. And remember, the Yankees' mission is to get to the World Series, which requires a seventh month of starts for three young pitchers who aren't ready for even six full months yet.
Without Santana, New York must plan for rotation instability in 2008. It can be done, but the odds begin to work against a team the more second-tier starters it has to plug in. ("Second tier" is not a blanket evaluation of talent -- sometimes a replacement is better than the original -- but a marker of stability.) Indeed, rotation stability has been one big reason why Boston has been winning the titles that used to belong to New York. This chart (above, right) offers a quick look at the number of starters used by AL East teams from 2004 through '07.
The Yankees have needed 10 more starting pitchers over the past four seasons than Boston and are well outside the range of every other team in the division. That's only part of the story, though. Look at the starts needed from second-tier pitchers -- that is, all starts made by everyone other than the five most-used starters each year:
Games Started by Second-Tier Starters
Team 2004 2005 2006 2007 Total
Yankees 36 51 26 41 154
Red Sox 5 19 48 22 94
Here the difference in rotation stability is even more apparent. Over the past four seasons the Yankees have handed the ball to second-tier starting pitchers 60 more times than did Boston. The Red Sox have done a better job identifying reliable starting pitchers and, by a combination of luck and design, keeping them healthy.
O.K., so what? How important is that? The Yankees were 23-18 in those 41 second-tier starts last season. And every team needs depth, right? After all, the average team uses 10 starting pitchers per year. But each of the past six world champions have been below that average, while the Yankees have been worse than average every year since their last World Series appearance, in 2003, when they needed only nine. (Since then New York has used 12, 14, 12 and 14 starters. It's the equivalent of a golfer having to scramble often to save par; it can be done, but with a higher degree of difficulty.)
Again, let's go back to the Schilling Theory, or its opposite corollary: keep the ball out of the hands of second-tier pitchers as much as you can. The more you don't have to rely on those No. 6 through No. 12 starters, the better off you'll be. The 2007 Cubs were another good example of this theory at work. They cut their second-tier starts from 54 to 10, helping them advance from 66 wins to 85 and from last place to first.
Further, let's look at the 12 full-season world champions of the wild-card era and examine how many times they had to dip into the second-tier pool of pitchers.
Starting Pitching Stability of Recent World Series Winners
Team Total Starters Second-Tier Games Started
1996 Yankees 12 28
1997 Marlins 8 30
1998 Yankees 10 20
1999 Yankees 9 10
2000 Yankees 12 24
2001 D-Backs 11 36
2002 Angels 8 22
2003 Marlins 9 19
2004 Red Sox 8 5
2005 White Sox 6 10
2006 Cardinals 9 30
2007 Red Sox 9 22
What you find is plenty of rotation stability. What jumps out is the remarkable reliability of the 1999 Yankees, 2004 Red Sox and 2005 White Sox, all of whom covered at least 94 percent of the season with their top five starters. Also, the 2001 Diamondbacks are the only team on the list with more than 30 second-tier starts, and of course, they were top-heavy in the rotation with Schilling and Randy Johnson. Looks like Schilling is on to something here.
Finally, one last word: I figured that teams burn through more starting pitchers today than ever before. After all, we hear about scores of injuries, the switch from a four-man rotation to a five-man rotation, kids rushed to the majors before they're ready . . . the usual stuff associated with the "watering down" of pitching. Turns out that is not the case. The number of starters used in a big league season has remained fairly flat for a decade: between 283 and 308 every year except 2005, when it actually sunk to 271.
Average Starter Performance
Season Starters GS/Starter
2007 10.1 16.1
1997 9.4 16.1
1987 9.8 16.5
1977 9.8 16.5
1967 11.0 14.8
1957 11.1 13.9
When you look at the bigger picture, it turns out that teams ran through more starting pitchers 50 years ago -- back when we like to think of those dudes being much tougher. Perhaps it can be explained by pitchers having less defined roles back then, and maybe by doubleheaders, too. This chart (right) provides snapshots, in 10-year increments, of the average number of starters used per team and the average number of starts made by those pitchers. Turns out that today's starting pitchers don't measure up too badly when compared to their historical counterparts.